The New Sudoku Players' Forum

[lunababy_moonchild]

I just wondered : given that the program generates a unique puzzle every time the new puzzle button is pressed, and thus the user never gets the same puzzle twice, does that mean that the puzzles my (copy of the) program generates are totally different to everybody else's ?

That is, are the puzzles generated by each copy of the software completely unique to that particular copy of the software or would my copy of the software eventually produce the same puzzle as someone else's copy of the software?

Luna

[ravel]

In my programs this is decided by the question, if the random seed (start value for the random function) is fixed or e.g. bound to the date/time. To check it, reinstall it twice and compare the first puzzles.

[ab]

Most likely the puzzles you get are different from ones someone else gets. Of course it's all random, so there's a very small chance you get a puzzle that someone else does, but you're probably more likely to win the lottery two weeks running.

[tarek]

I'm with ravel,

Open you program for the first time, generate an easy puzzle & register the result.

Close the program.

Open the program again, genrate an easy puzzle.....if the results match then the puzzles are not "THAT" random.

tarek

[Son of Pappocom]

Hi luna,

Ravel mentioned the seed. Computers can't really generate true random numbers, so a seed number is used (something like milliseconds elapsed since 12:00am, January 1, 1970, or somesuch) to generate a pseudo-random number. This increases the likelihood that a "random" process will be repeated, but it's still a very long shot!

Regards
Scott

[lunababy_moonchild]

Thanks everybody for your answers, I was just wondering if everybody got different puzzles and pondering the magnitude of that!

Luna

[wapati]

Thanks everybody for your answers, I was just wondering if everybody got different puzzles and pondering the magnitude of that!

Luna

Think of lottery tickets. Thay don't store all the numbers and give you one based on the last one sold. Every machine generates a number (numbers) and then reports what was sold.

Making new ones is easier than running a database, if large numbers are involved.

[lunababy_moonchild]

Think of lottery tickets. Thay don't store all the numbers and give you one based on the last one sold. Every machine generates a number (numbers) and then reports what was sold.

Making new ones is easier than running a database, if large numbers are involved.

Yes, I know but the lottery is one machine and it does it once a week (originally, and in Gt. Britain). My question was about each copy of Pappocom's Software of which there were numerous copies sold and I wondered if each copy created entirely unique puzzles, which were unique to that copy of the program or if your copy (say) created the same puzzles mine did.

Luna

[wapati]

I was meaning the VENDING machines.

[daj95376]

Most prime number generators use two prime (or relatively-prime) numbers and a process (that I believe is) called multiplicative congruence to create a cyclic chain of pseudo-random numbers. A seed value is used to select an initial value from this chain. A common practice is to use the milliseconds elapsed value (mentioned earlier) for the starting seed when the program starts ... or when a new puzzle is generated.

The important thing on differentiating your puzzles from someone else's puzzles is where each of you enter the chain. In the old days of Microsoft FreeCell, the chain was only 32K in length and that limited the number of unique games -- which were identified by their starting seed number. Today, most random number generators use a full 32-bits for the seed ... and this leads to over 4,000,000,000 entries in the chain. Your chances of having the same puzzle as someone else is very slim!

[lunababy_moonchild]

I was meaning the VENDING machines.

My mistake, I was thinking of the lottery draw on TV

Most prime number generators use two prime (or relatively-prime) numbers and a process (that I believe is) called multiplicative congruence to create a cyclic chain of pseudo-random numbers. A seed value is used to select an initial value from this chain. A common practice is to use the milliseconds elapsed value (mentioned earlier) for the starting seed when the program starts ... or when a new puzzle is generated.

The important thing on differentiating your puzzles from someone else's puzzles is where each of you enter the chain. In the old days of Microsoft FreeCell, the chain was only 32K in length and that limited the number of unique games -- which were identified by their starting seed number. Today, most random number generators use a full 32-bits for the seed ... and this leads to over 4,000,000,000 entries in the chain. Your chances of having the same puzzle as someone else is very slim!

Ah hah, with you now. It's miraculous isn't it? Absolutely brilliant, I wish I'd thought of that!

Luna

*edit : spelling mistake*

[Smythe Dakota]

.... Your chances of having the same puzzle as someone else is very slim!

And (to address Lunababy's initial confusion) this slim chance of having the same puzzle twice would neither increase nor decrease if you used two copies of the program, rather than using the same copy twice.

Bill Smythe